Dear Comp-neuro community,

I am following up on my earlier PSI announcement with a new perspective that corrects its clinical framing and proposes a staged validation protocol:

Reframing the Psychiatric State Index Through Clinical Assessment and Measurement Validation
Full manuscript with references

The manuscript is not peer reviewed. Psychiatric review has been requested and remains pending. The computational project currently provides numerical signal analyses and exploratory diagnosis-group benchmarks; its available data do not validate future clinical-event prediction.

The protocol makes clinical adjudication the reference standard and asks whether EEG adds value over a clinical-only comparator. It separates synchrony/coupling, diagnosis-group differences, movement-associated measurements and future-event prediction. Scalp gamma/PAC is excluded from primary human analyses, with independent artifact measurements required for secondary interpretation; intracranial LFP is treated separately.

The accompanying PR #3 corrects PAC surrogate boundaries and configured-count reporting, enforces acquisition/QC compatibility, and balances human training by participant. Synthetic regression tests and public-data benchmarks are documented as computational evidence, not clinical validation.

I would welcome criticism of the surrogate/null assumptions, participant-level evaluation and incremental-value study design, particularly from researchers with psychiatric or EEG measurement expertise.

Best regards,
Netanel Stern